SGU Episode 865: Difference between revisions

From SGUTranscripts
Jump to navigation Jump to search
m (added Rogues)
(1st news item transcibed)
Line 393: Line 393:


== News Items ==
== News Items ==
'''S:'''
'''B:'''
'''C:'''
'''J:'''
'''E:'''
<!-- ** the triple quotes are how you get the initials to be bolded. Remember to use double quotes with parentheses for non-speech sounds like (laughter) and (applause). It's a good practice to use brackets for comments like [inaudible] and [sarcasm]. -->
''(laughs)''
''(laughter)''
''(applause)''
[inaudible]


=== Machine Learning and Mental Health <small>(16:32)</small> ===
=== Machine Learning and Mental Health <small>(16:32)</small> ===
Line 424: Line 408:
-->
-->
* [https://news.mit.edu/2022/deploying-machine-learning-improve-mental-health-rosalind-picard-0126 Deploying machine learning to improve mental health]<ref>[https://news.mit.edu/2022/deploying-machine-learning-improve-mental-health-rosalind-picard-0126 MIT News: Deploying machine learning to improve mental health]</ref>
* [https://news.mit.edu/2022/deploying-machine-learning-improve-mental-health-rosalind-picard-0126 Deploying machine learning to improve mental health]<ref>[https://news.mit.edu/2022/deploying-machine-learning-improve-mental-health-rosalind-picard-0126 MIT News: Deploying machine learning to improve mental health]</ref>
'''S:''' All right well Cara you're going to start us off with the use of machine learning and mental health.
'''C:''' Yes and spoiler alert─
'''B:''' Got my attention.
'''C:''' ─someone named Dr Steven Novella wrote about this today on [https://sciencebasedmedicine.org/using-ai-technology-to-treat-mental-illness/ Science-Based Medicine]
'''B:''' Oh he's a punk.
'''C:''' I really like this story because it's not really about the background, I'll give you the background, it'll be quick and dirty. But really it's about it's, it's about the conversation that it opens up. So long story short a researcher from MIT and a researcher from Master General one of whom is a computer scientist or really an AI researcher who focuses on machine learning. The other of whom is a mental health researcher, you know behavioral health. Decided to join forces and try and see what happens when we collect data about mental illness and specifically they were looking at depressive disorders. Let's see what happens when we collect a whole bunch of data, about depressive disorders and feed it into a machine learning algorithm. And so we know that machine learning is like a kind of a sub type of, AI in which, and we talk about machine learning on the show all the time. We've actually dug really deep into the mechanisms of machine learning before on the show. Although it is to some extent still a black box. But the idea here is if you feed a machine learning program a bunch of information, like real world data it will then be able to categorize sort and start to make determinations based on that data. And like I said it's a bit of a black box we don't really know what the algorithm they're using are because it it's iterative. Or I should say recursive?
'''S:''' It's, it's both but yeah yeah, it was recursive because the other layer I would add there for this kind of machine learning is that there had the the program has to make some kind of determination.
'''C:''' Yeah.
'''S:''' And and then gets feedback, like it knows if it's correct or not. And that's.
'''C:''' Right right there has to be an external sort of validation to that.
'''S:''' Yeah, there's got to be some feedback, yeah.
'''C:''' Yeah and so the idea here is okay what are the determinants of a depression diagnosis. So some self-report stuff some assessment information and then also like what are some behaviors that are associated with depression. So they were looking at things like wearables which I mean we're going to talk about that in a second but they're looking at things like heart rate and skin conductance and blah blah and sleep and and activity level. But they were also like data mining with texts which I think is probably honestly the number one piece of data that's the most predictive. And and the self-report surveys of the individual saying you know this is how I feel you know let me rate my symptoms x to y. And then the idea is with external validation they wanted to first say okay does this comport with clinical diagnoses. I think that's like in in many ways the first step, right? Like does this work? Well we have to know and we need to externally validate it well what's the way to do that if these people sit down with a clinician is that clinician going to diagnose them with you know x disorder. People love to talk about this in a really saccharine way and say oh it could be great, it could be you know, it could blow the cover off of access to mental health treatment. And to some extent I feel very very hopeful about the democratization of the sort of "new self-help movement" I think because the "new self-help movement" in mental health is much more evidence-based. But there's always room for for quack technology and and and quack not technology but quack utilizations of the technology. And when I make that statement I want to caveat it. We are living in a culture and a society that is more obviously, has more access to to tools online than ever before. And a lot of mental health treatment is actually psycho-education. It's about learning about what a disorder is learning cognitive behavioral interventions tools and tricks to help understand our thought disturbances, and our thought distortions. And help correct some of these things that's not all of mental health treatment but it's an important part. And so whereas in the past, and still to a huge extent today, sadly people would think oh I want to better myself I'm going to read some like {{w|Tony Robbins}}. Now people do have access to reading legitimate literature about mental illness and they have access to evidence-based apps and evidence-based workbooks. Manuals that are used within mental health treatment. The problem is they're not guided by a clinician.
I think there's multiple things that are important to tease out here. The first one being and it's it's something that I don't think has been pointed too much in in the coverage of this, is that there's a big difference between having a diagnosable mental illness and experiencing "typical symptoms". Like there's a difference between having major depressive disorder and sometimes being depressed. And that difference is somewhat arbitrary because it's a construct right? It was invented by people but those people have all kind of come together to agree and publish standards. Those standards are constantly in flux but for the most part they publish standards about you know, these are the the sort of qualifications for somebody to have a diagnosable mental illness and maybe they need medication management or therapeutic management. Versus you know sometimes I get sad. And I think there's also a difference in how you approach the sometimes I get sads with the this is clinical depression that is interfering with this person's ability to function and potentially there are some like concerns around suicidality or concerns around you know really terrible dangerous outcomes. And and I think making that distinction is important when we talk about the use of these technologies in the future. And I think sometimes sadly they get lost in the shuffle. People love to talk about oh maybe you could send an alert you know their smart watch could send an alert to their physician to tell them that they're potentially suicidal. And I'm like first of all I don't I don't know if that's ever gonna happen. Maybe. But second of all that's a very different utilization of this technology than indicating to you you're feeling low maybe it's time to go on a walk.
'''S:''' Yeah there's definitely different levels there but it's sort of all the same idea that it's going to gather data, look for patterns, correlating your behaviors with your mood either self-reported or evaluated by your physician. And then when it starts to see the same pattern of behaviors happening again it tells you hey you're slipping into the same pattern maybe you should sleep more. Or get some more physical exercise get out of the house you know or whatever. But that is would would have to that's beyond the scope of the development of this technology.
'''C:''' For sure that's all in the discussion section.
'''S:''' Yeah now you're getting into clinical trials of applying this technology. Then again when I wrote about it as of my main fear is I mean the technology will work it will do what it's supposed to do. I mean these AI algorithms are way more powerful than need to be for something like this.
'''C:''' Yeah we know this from marketing data right like yeah like Facebook knows more about me in some, in some ways than I know about myself.
'''B:''' Yeah.
'''C:''' Like they know what I want to buy they know what I'm gonna click on just based on my patterns that I've never taken the time to introspect about and that's sort of what we're talking about.
'''S:''' That's kind of a deeper question of how predictable are humans. We like to think that we're all individual snowflakes ''(laughter)'' but psychological research is kind of based on the idea that people kind of do the same thing. At least 60-70% you know what I mean, like it's not never 100% but you can.
'''B:''' It's not a coin flip but yeah.
'''E:''' There are patterns that can be determined sure.
'''C:''' Yeah and a social social psychologist, you know think think {{w|Daniel Kahneman|Kahneman}} think these different researchers who have you know kind of joined this idea of economics with psychology. Yeah we can predict behavior pretty pretty well. You're right some people are going to be on the tails of that normal curve but, yeah, but there's a normal curve.
'''S:''' Yeah, but there's a normal curve. We're not all over the place. {{w|Richard Wiseman}} talks about this all the time that's what he researches his interest is in looking at how to predict how people will behave with very specific stimuli.
'''E:''' Isn't that what like mentalism is practically about?
'''S:'' Yes, yes absolutely.
'''C:''' For sure.
'''S:''' Mentalism wouldn't work if people were unpredictable, magic wouldn't work if people weren't predictable you know absolutely. So it's partly based on that premise, that people are going to be fairly predictable although this could theoretically also individualize the basic pattern you know, it's looking for your patterns not just the patterns of depression.
'''C:''' Right.
'''S:''' But then figuring out how to use, it's going to be the tricky part. And you know it could be they they talked about in the article recommending changes but that could be flipping the direction of cause and effect. It's like you're not sleeping as much and you're depressed so sleep more. Or it could just be that not sleeping is not causing you to be depressed it's a result of you being depressed. Or they're both the result of something else happening like increased stress and just saying sleep more is like a facelift it doesn't really treat the underlying problem which is changing a marker so this is where this is where the clinical research has to come in.
'''C:''' Yeah. I think in some ways that's one of the biggest problems with {{w|Cognitive behavioral therapy|CBT}}, I think that you know we have made this the gold standard because it easily can be utilized in randomized controlled trials it lends itself to scientific. Like the way that we test drugs we apply to these psychological interventions and CBT just is good at that because it's highly manualizable, it's highly consistent, you can sort of do the same intervention to different groups that are randomized. The problem is that very often what we're looking at is a very acute intervention for a very acute symptom. And we're not really looking necessarily at why is that person having the panic attack. Like we might be able to intervene and teach them really good tools to sort of get grounded again, to get their breath back and to feel calm and feel safe and sort of not not go into the throes of panic. But what we're not getting deep into is why panic attacks consistently happen in their lives.
'''S:''' Yeah and that's I don't think this tool is for that you know.
'''C:''' That's what mental health treatment is for.
'''S:''' I agree, I think this will this is obviously not going to replace mental health treatment this is one more tool that could be useful for gathering behavioral information about specific individuals. And then also maybe providing some useful feedback either to them or to their practitioner. And you know you could you can envision a lot of situations where somebody who has obsessive-compulsive disorder they're starting to get into more and more obsessive behavior. Or someone who's anxious, who's starting to get physiological signs of anxiety that's starting to ramp up.
'''C:''' That they may not now this right away.
'''S:''' Oh yeah or somebody with depression or yeah leaving the house as much as they did yeah whatever. They might not be consciously aware of it yeah until it gets more severe and it might just trigger and sort of a more of an early warning system. So it'll be useful I think. But but at the same time I mean I think it will be used appropriately. But it will also then be completely abused it will be like─
'''C:''' Absolutely.
'''S:''' Download this app and treat your depression with this individualized personalized AI powered algorithm that will find out what what's making you depressed, you know what I mean. It'll be─
'''C:''' Yeah. Traditional therapy is hard and expensive but you can do this in only five minutes a day.
'''J:''' While you're sleeping.
'''E:''' So it'll unfairly target those who can't afford otherwise, to get the right kind of.
'''S:''' Well I don't know I don't know about that I, because this these kinds of things tend to target the worried well people who have disposable income. In fact there was a recent study that's that showed that the advent of tele-health you know with the pandemic, decreased socioeconomic disparities in access to healthcare.
'''E:''' Cool.
'''C:''' For sure it did.
'''S:''' So it's good it's been a very positive thing.
'''C:''' Yeah and I think an app could do that too.
'''S:''' Sure.
'''C:''' Because I think what a lot of people, what's important to realize if you've never gone through a heavy course of mental health treatment or sort of peeled the curtain back behind this whole industry, is that you know I had a supervisor who said it really well recently. She always tells her patients at the beginning my goal is to make myself...
'''S:''' Irrelevant, redundant.
'''C:''' Yes thank you not irrelevant or redundant but like my goal is to make─
'''S:''' Unnecessary.
'''C:''' ─unnecessary, my goal is to help you become your own therapist.
'''S:''' Yeah and that goes beyond mental health Cara because, especially with the advent of smartphones and technology you know a lot of modern medicine is teaching patients to take care of themselves.
'''C:''' Yeah it's like preventive care.
'''S:''' Yeah. Like yeah you can't be looking over your shoulder. You're gonna treat your own diabetes or migraines or whatever it is I'm just going to give you the tools to do it and teach you how to do it, but you've got to do it day-to-day because you're the one living with yourself you know I can't be there─
'''C:''' Absolutely.
'''S:''' To monitor what you're eating and you give you your insulin or whatever you got to do that. So this that's these kind of apps I think will help with that paradigm where we are really helping patients manage themselves. Because now this becomes an extra tool.
'''E:''' Patients generally do bad poor managers of their own when left to their own devices?
'''S:''' They're all over the place there but there are patients who are excellent, who are you know just like almost too much, like they're so completely on to you know on top of things and at the other end of the spectrum there are people like they don't even know what medications they're taking.
'''C:''' Yeah.
'''S:''' You will see the full full spectrum.
'''C:''' This tool could be used for very nefarious purposes or incredibly beneficial purposes.
'''S:''' Yeah and it will be used for both.
'''C:''' And it will be used for both yeah yeah.
'''S:''' Totally.
   
   
=== Hardy Tardigrades <small>(30:17)</small> ===
=== Hardy Tardigrades <small>(30:17)</small> ===

Revision as of 22:06, 5 March 2022

  Emblem-pen.png This episode is in the middle of being transcribed by Hearmepurr (talk) as of 2022-03-05.
To help avoid duplication, please do not transcribe this episode while this message is displayed.
  Emblem-pen-green.png This transcript is not finished. Please help us finish it!
Add a Transcribing template to the top of this transcript before you start so that we don't duplicate your efforts.

Template:Editing required (w/links) You can use this outline to help structure the transcription. Click "Edit" above to begin.


SGU Episode 865
February 5th 2022
865 Tesla Bot.jpg
(brief caption for the episode icon)

SGU 864                      SGU 866

Skeptical Rogues
S: Steven Novella

B: Bob Novella

C: Cara Santa Maria

J: Jay Novella

E: Evan Bernstein

Quote of the Week

Officials, educators, media personalities, and opinion makers must do everything they can to educate the public, assuage people’s fears, and show that the pandemic is being handled fairly and rationally. Lives and people's well-being are at stake, for years to come.

Nidhal Guessoum, Algerian astrophysicist

Links
Download Podcast
Show Notes
Forum Discussion

Introduction

Voice-over: You're listening to the Skeptics' Guide to the Universe, your escape to reality.

S: Hello and welcome to the Skeptics' Guide to the Universe. Today is Wednesday, February 2nd, 2022 and this is your host, Steven Novella. Joining me this week are Bob Novella...

B: Hey, everybody!

S: Cara Santa Maria...

C: Howdy.

S: Jay Novella...

J: Hey guys.

S: ...and Evan Bernstein.

E: Well it's Groundhog's Day again.

S: Groundhog's Day? Yep it's 2 two two zero two two.

E: And numerologists heads are exploding all over the planet.

S: Yeah February 22nd will be the next date where it's all 2s and then I think that's the last time that happens this century, right?

E: Until the year 2202, so 20 000 years from now.

J: I didn't see any weirdness about it which was good, you know you figured somebody would have said something stupid.

S: Yeah it's true I even didn't, didn't know it was Groundhog's Day until Evan mentioned it right before.

E: You forgot?

C: So I have to tell you guys something fun that happened the other day.

E: Oh OK.

C: And by the other day I actually mean it happened in December and I've been trying to remember to tell you every week and every week I forget. So I'm very excited now. This past December I went to Vegas with a friend who is a physician and we attended a conference that was put on by the City of Hope here in Los Angeles about end-of-life stuff. Also went and gambled in between sessions and by the way here's my advice if you got to stay in Vegas stay in a hotel that doesn't have a casino it was so nice.

E: They have those?

C: Yeah the Waldorf doesn't have, I mean it's too expensive the only reason I was there because we had block deals for the conference but yeah no casino so no cigarette smoke no nothing like very clean, very nice.

B: Also no gambling.

C: No noise yeah but you just walk to the casino next door.

B: That's true.

C: Leave that behind. Anyway so we decided to go someplace off strip and we, we called it a Lyft and the Lyft picked us up we're in the lift we're talking talking talking in the back seat like you do, fully masked you know because this is height of the pandemic and the Lyft driver goes "Are you Cara Santa Maria?" and I was like yeees.

E: Maybe.

C: And he was like "I listen to you all the time on the Skeptic's Guide to the Universe" (laughs) and he recognized me by my voice because─

B: Wow.

C: I know he couldn't have seen my face I was like wearing a black mask and a black, it was cold too so I was wearing like a beanie and a─

E: All bundled up, yeah.

C: ─puffy jacket and stuff yeah and he─

B: Wow.

C: He could hear my voice and.

B: Is that what he said tell me all about Bob?

C: (laughs) He did I forgot about that Bob yes he did he did, he asked for a lock of your hair.

B: That's amazing though you're just your voice.

J: Did it creep you out at all?

C: No I loved it the guy was so sweet I asked him like what Vegas was like, he told me his whole story, he definitely was working in a in a kind of skeptic-y kind of a field you know, was kind of answering all my questions about what it's like being a local in Vegas, that must be kind of an interesting phenomenon right? Vegas is one of those cool cities where the population is relatively low but the floating population is very high.

S: Yeah the tourist population's high.

C: It's huge yeah and so just kind of what it's like being an insider in Vegas and everything and he was super sweet so shout out to Jonathan my Lyft driver.

E: Hey Joe.

C: Hopefully you're listening right now. You were very nice.

J: He's gonna drive off the road Cara.

(laughter)

S: We've all been recognized you know a couple of times, out in the wild you know, like not associated with a skeptical event or something where people are coming to see us but just like in the general po, gen pop you know and it's usually my voice.

B: Gen pop (laughs)

C: It's weird when it's by voice.

S: Because you think about it, you're you're more likely to attract attention that way, you know people will hear you know you hear a voice you recognize in an airport or a lobby or whatever.

B: Or a horror convention.

S: Yeah.

E: Especially when our medium is an audio based medium so.

C: But it's weird that sometimes they they say a lot like I didn't know what you looked like, so so for the vocal recognition to overpower any lack of visual recognition, is just that much more powerful.

E: Oh I could easily identify people if they were in the room just from their if they were radio personalities from you know─

C: Really?

E: ─from past 20 or 30 years absolutely I feel I could recognize them and pick them out of a crowd.

S: But think how much more weird that must be for them than for us right, so imagine you're a Lyft driver and you hear a voice you recognize. Yeah that that's, it's a much less likely event for them.

E: Right.

S: I so the first time that happened was in an airport you know, and it wasn't associated with it, it was like a vacation or something, there was no skeptical event and just some random person walked up to say "Are you Stephen Novella?". So I calculated what the odds of that would be, and it was it was very high because you think about the number of listeners we have versus the number of people in the country. You know it's order of magnitude about every 3 000 people you know there should be somebody in there who listens to us. And so I was in an airport with you know several thousand people someone in there probably was a listener of the show. So it wasn't that that big of a coincidence you know.

C: But being paired up with a with a driver.

S: That driver's a little bit more yeah.

C: Yeah that's that's funny.

S: Are you a, Cara you are but are you happy that this is a rare event or like how?

C: Oh for sure. I there there have been times in my life where I got recognized more often when I was on like bigger tv shows and actively on them. And I mean it never was huge I was always very lucky it was like minimal. And there have been times when I've been recognized for you know the podcast and for other work that I've done. But of course I've dated people and I've had dear friends in my life who have quite a bit of success and it's annoying to them. Like I feel kind of bad for them. I think there's a part of it that's exciting but there's also a part of it where they struggle to have privacy. Like they have to think about what they look like every time they leave the house. And they have to you know there have definitely been times when I've been out at dinner with somebody and people just walk up to the table and start talking to them and they're like "thank you so much but like i'm on a date".

J: Yeah sure.

B: Oh yeah.

C: You know can you give us a privacy?

E: Yeah.

S: Right.

C: That must be tough you know and even those people would say I'm lucky I don't have paparazzi waiting in my driveway for me to leave my house.

S: Yeah, there's different levels of fame, yeah.

E: That's right.

B: Yeah the upper levels.

C: Like I just get a good table at every restaurant but I don't have people like you know hounding me all the time.

B: Yeah there are certain levels that must be like totally life-changing, absolutely life-changing whether you can't go out in public and you you know you always have to think of you know boy what's going to happen today with the paparazzi. What are you, what are they going to you know they're going to catch me picking my nose in public again? Or or something worse far worse these poor people that are caught you know, getting out of a car I mean this is like oh come on really?

C: But not all of them are that's the other sad thing. Some people are famous because of something that happened to them because of a court case they were involved in right because.

S: Yeah this is you know part of a bigger issue of just privacy in general, just for regular citizens you know let alone you know the people who are famous. And we may have to figure out ways to carve out islands of privacy you know in the otherwise social media you know constant you know life where we at any point somebody could be filming us or taking our picture recording us or whatever.

B: It's just going to get, it's just going to get worse.

S: It's going to get worse.

B: Yeah imagine when there if there's when there's you know quintillions of particles of smart dust floating─

E: Oh forget it. Right, exactly.

B: ─everywhere each one is like a high-res video camera it could be anywhere.

C: Oh god stop.

E: You'll be breathing cameras into yourself.

C: You're so good at describing the most like dystopic scenarios.

B: Oh I got lots of them here's Bob's you-ula.

J: Cara, Bob likes it.

C: I know. It's utopic to him.

E: It's the Brave New World, right?

B: Yeah, yeah the tech is awesome but some of the downstream effects I'm sure won't be. But many of them will be, double-edged sword.

Quickie with Bob (7:57)

S: All right Bob you're actually going to start us off with a Quickie with Bob, and this Quickie I'm going to tell you what it's about because you stole my one of my science or fiction items.

E: Good, we'll only have two items to guess from tonight.

S: And you just tell me, tell me until right at the beginning of the show you were, you were late in telling me your topic for this so you're gonna tell us.

B: So now you're gonna...

S: I'm gonna, I'm gonna steal your thunder I'm gonna say you are gonna tell us about scientists who basically invented flubber.

E: Fiction.

B: Yes thank you Steve I knew you had your eye on me this this is your Quickie with Bob now gird your loins this one is cool.

C: Ugh.

B: That's an expression I don't know. That's a great expression.

C: What does gird mean, what does that verb mean?

E: You look at your loins and you're like grrrrd.

B: It's only you know it's only used in the context of gird your loins as far as I can tell.

C: I'm looking it up.

B: And I love it, please do. Okay while you do that, so teams a team of researchers from the university of Massachusetts Amherst have developed a new meta material, my third favorite material in the world, with properties not seen in nature. Which is kind of what meta materials are. So researchers compare it to a rubber band for example, a rubber band that you know, you can fling it across the room right? You pull it back you let go and you fling it across the room. But this one if you go with this analogy, this one if you stretch it past a certain point before letting it go it flies for a mile. So that gives you an example of kind of what they're talking about in a way. It's an elastic material with magnets embedded in it. It's elastomagnetic material that harnesses the energy released by phase shifts. Now think back 11th</supe>, 11<supe>th grade physics. Phase shifts like when liquid turns into a gas, water turns into steam, releases energy right, there's a lot of energy that can be released in these types of phase shifts. Now there are some phase shifts though that happen when you go from one solid phase to another solid phase and but you still have this energy that's released. Now this phase energy if you will can be released or absorbed and the magnets can control when that happens. So what this could do then is this met- this elastomeric meta material can do then is allow a large amount of energy to be absorbed or it can be released quickly, very quickly for an explosive movement. So yes this is the path to a black panther suit. Clearly, because his suit if you remember the movie had two of those, had both of those properties but I'm sure they were you know well whatever. So until until that happens we may see it in robots to give them more power─

E: Oh yeah.

B: ─using it without using additional energy, conventional energy anyway. We'll see it in helmets and other protective materials which would be able to dissipate a lot of energy very quickly. So I have to keep this under a minute which I think I blew past. So now Cara loins ungirded this has been your Quickie with Bob I hope it was good for.

C: Bleh, it took on a whole new meeting with the girding and ungirding.

B: Yes it it seems especially good for this Quickie with Bob so I may keep it.

E: So the magnets retain the energy and you it can be released upon a trigger happening is that?

S: No the magnets are more about controlling the the the phase shift, the phase shift is what stores the energy so a phase shift is like going from liquid to ice you know but it could also be one state of matter to another state of matter and if one and if one of those states one of the configurations requires is a much greater higher energy state when you shift from the low energy state to the high energy state you absorb all that energy that's now being stored in the higher energy state. And if you go back it releases all of that energy so it's a you know it's a rapid energy storing system but that also because it's like a rubbery material it could be used as Bob said for armor or the soles of your shield so that you could dunk baskets.

B: Meta materials are endlessly fascinating because when because when you when you describe a material by invariably, this is a material that is not found in nature I mean it is do when you pattern things at the at the nano scale like like this like like they're doing with typical meta materials, it is amazing what they're accomplishing and I, I cannot wait to see you know what what other new materials are are coming out. It's it's like never ending there's like one new meta material after another. I've read about another one just today that the description of it was electro momentum, look it up, I mean it's just so fascinating and this is one that I just wouldn't have even imagined electro you know elastomagnetic meta material. Incredible. And and harnessing phase that you know the energy released by a phase change, my god, incredible. I think we're just scratching the surface with just with this specific meta material. Who knows where this could go.

Random Asides about English Words (12:44)

S: Bob I want you to imagine doing the following things let's say you're wearing a tunic that goes down to your ankles. You pick up the bottom of your tunic pull all the fabric above your knees.

B: Yep.

S: Now gather all the material in front of you right. So it's tight around your bottom but all the extra material is in front of you. Bring that material down between your legs to your backside.

B: Gotcha.

S: Then you grab each you know sort of half of the material bring it around to the front and tie a knot in it.

B: OK.

S: You have now girded your loins.

C: (laughs)

B: Nice.

C: Exactly, yeah, to encircle or bind and secure and yeah it has translated into we're girding for battle we're getting ready to action.

S: Yeah you couldn't fight or work with tunic going down to your ankles you had to gird your loins before you could go into battle.

C: And of course yeah what about the word girdle, yeah similar.

B: There you go there you go.

C: Also such a gross word. Girdle.

S: My girdle.

E: Girdle the turtle.

S: Girdle the turtle.

B: Is there, Steve and Cara, I haven't looked it up is there another give me a use of the word gird that that's you know that's viable.

C: Yeah I need to gird my sword to secure it with yeah.

B: It makes sense but nobody says it I, I think it's one of those words that really only remains in this one expression in English, gird your loins, otherwise.

S: The first definition is encircle a person or part of the body with a belt or band.

C: But interestingly Merriam-Webster actually lists that as the second definition and their first definition is to prepare yourself for action.

B: Yes that's it.

C: So they sort of seeing this as sort of Bob's way.

B: Yeah that's one of those words that just it's one of those words that when you think about it you're like yeah, I know this word it's in this expression I know it very well I've known it for decades. But that word is not used anywhere else. It's a remnant.

S: So like fell, like one fell swoop do you ever use the term fell as in one fell swoop in any other context?

B: Yeah.

S: You could.

C: Or like or like home in on or like a homing pigeon those are the only references you ever use the word home. But I guess it's more specific there I don't know. My favorite is once I had a friend ask me Cara define the word such and I was just stumped, couldn't do it.

J: I can't do it.

E: Such...

B: Wow it's a good one.

C: Yeah.

S: Yeah.

E: What part of speech is "such".

S: It's a and such.

J: Like similar to it was like.

C: The first definition in Merriam-Webster is "of a character of a kind or character to be indicated or suggested", what?

S: Yeah it's a weird, it's a weird one.

B: That must be a very hard word to learn, when you're learning English, how to use that word properly.

E: Or pronounce it, they all say suck.

C: I bet you interestingly I bet you a lot of people for whom English is not their first language unless they really are fluent they probably just don't use it because it's not really necessary.

S: Well I mean but there's something like if you say oh he is such a jerk how would you say that without?

C: He's a total jerk.

S: He's a yeah.

C: Yeah he is the jerkiest of all jerks. I think there are ways to work around it.

S: Yeah yeah.

B: There's a lot of ways, to of ways.

C: It's kind of like in Spanish if you don't know the if you can't conjugate in the future tense, you can just say "voy a" like "I'm going to". Just put that before everything and it's a crutch─

B: Voy a hablar conmigo.

C: ─and you see that a lot with people for whom English is not their first language they might use the same tense over and over because it's like why do I need to learn all these tenses I can just speak in the present tense.

S: I'm liking this very much.

C: Exactly, yeah right that's a function of language learning.

News Items

Machine Learning and Mental Health (16:32)

S: All right well Cara you're going to start us off with the use of machine learning and mental health.

C: Yes and spoiler alert─

B: Got my attention.

C: ─someone named Dr Steven Novella wrote about this today on Science-Based Medicine

B: Oh he's a punk.

C: I really like this story because it's not really about the background, I'll give you the background, it'll be quick and dirty. But really it's about it's, it's about the conversation that it opens up. So long story short a researcher from MIT and a researcher from Master General one of whom is a computer scientist or really an AI researcher who focuses on machine learning. The other of whom is a mental health researcher, you know behavioral health. Decided to join forces and try and see what happens when we collect data about mental illness and specifically they were looking at depressive disorders. Let's see what happens when we collect a whole bunch of data, about depressive disorders and feed it into a machine learning algorithm. And so we know that machine learning is like a kind of a sub type of, AI in which, and we talk about machine learning on the show all the time. We've actually dug really deep into the mechanisms of machine learning before on the show. Although it is to some extent still a black box. But the idea here is if you feed a machine learning program a bunch of information, like real world data it will then be able to categorize sort and start to make determinations based on that data. And like I said it's a bit of a black box we don't really know what the algorithm they're using are because it it's iterative. Or I should say recursive?

S: It's, it's both but yeah yeah, it was recursive because the other layer I would add there for this kind of machine learning is that there had the the program has to make some kind of determination.

C: Yeah.

S: And and then gets feedback, like it knows if it's correct or not. And that's.

C: Right right there has to be an external sort of validation to that.

S: Yeah, there's got to be some feedback, yeah.

C: Yeah and so the idea here is okay what are the determinants of a depression diagnosis. So some self-report stuff some assessment information and then also like what are some behaviors that are associated with depression. So they were looking at things like wearables which I mean we're going to talk about that in a second but they're looking at things like heart rate and skin conductance and blah blah and sleep and and activity level. But they were also like data mining with texts which I think is probably honestly the number one piece of data that's the most predictive. And and the self-report surveys of the individual saying you know this is how I feel you know let me rate my symptoms x to y. And then the idea is with external validation they wanted to first say okay does this comport with clinical diagnoses. I think that's like in in many ways the first step, right? Like does this work? Well we have to know and we need to externally validate it well what's the way to do that if these people sit down with a clinician is that clinician going to diagnose them with you know x disorder. People love to talk about this in a really saccharine way and say oh it could be great, it could be you know, it could blow the cover off of access to mental health treatment. And to some extent I feel very very hopeful about the democratization of the sort of "new self-help movement" I think because the "new self-help movement" in mental health is much more evidence-based. But there's always room for for quack technology and and and quack not technology but quack utilizations of the technology. And when I make that statement I want to caveat it. We are living in a culture and a society that is more obviously, has more access to to tools online than ever before. And a lot of mental health treatment is actually psycho-education. It's about learning about what a disorder is learning cognitive behavioral interventions tools and tricks to help understand our thought disturbances, and our thought distortions. And help correct some of these things that's not all of mental health treatment but it's an important part. And so whereas in the past, and still to a huge extent today, sadly people would think oh I want to better myself I'm going to read some like Tony Robbins. Now people do have access to reading legitimate literature about mental illness and they have access to evidence-based apps and evidence-based workbooks. Manuals that are used within mental health treatment. The problem is they're not guided by a clinician.

I think there's multiple things that are important to tease out here. The first one being and it's it's something that I don't think has been pointed too much in in the coverage of this, is that there's a big difference between having a diagnosable mental illness and experiencing "typical symptoms". Like there's a difference between having major depressive disorder and sometimes being depressed. And that difference is somewhat arbitrary because it's a construct right? It was invented by people but those people have all kind of come together to agree and publish standards. Those standards are constantly in flux but for the most part they publish standards about you know, these are the the sort of qualifications for somebody to have a diagnosable mental illness and maybe they need medication management or therapeutic management. Versus you know sometimes I get sad. And I think there's also a difference in how you approach the sometimes I get sads with the this is clinical depression that is interfering with this person's ability to function and potentially there are some like concerns around suicidality or concerns around you know really terrible dangerous outcomes. And and I think making that distinction is important when we talk about the use of these technologies in the future. And I think sometimes sadly they get lost in the shuffle. People love to talk about oh maybe you could send an alert you know their smart watch could send an alert to their physician to tell them that they're potentially suicidal. And I'm like first of all I don't I don't know if that's ever gonna happen. Maybe. But second of all that's a very different utilization of this technology than indicating to you you're feeling low maybe it's time to go on a walk.

S: Yeah there's definitely different levels there but it's sort of all the same idea that it's going to gather data, look for patterns, correlating your behaviors with your mood either self-reported or evaluated by your physician. And then when it starts to see the same pattern of behaviors happening again it tells you hey you're slipping into the same pattern maybe you should sleep more. Or get some more physical exercise get out of the house you know or whatever. But that is would would have to that's beyond the scope of the development of this technology.

C: For sure that's all in the discussion section.

S: Yeah now you're getting into clinical trials of applying this technology. Then again when I wrote about it as of my main fear is I mean the technology will work it will do what it's supposed to do. I mean these AI algorithms are way more powerful than need to be for something like this.

C: Yeah we know this from marketing data right like yeah like Facebook knows more about me in some, in some ways than I know about myself.

B: Yeah.

C: Like they know what I want to buy they know what I'm gonna click on just based on my patterns that I've never taken the time to introspect about and that's sort of what we're talking about.

S: That's kind of a deeper question of how predictable are humans. We like to think that we're all individual snowflakes (laughter) but psychological research is kind of based on the idea that people kind of do the same thing. At least 60-70% you know what I mean, like it's not never 100% but you can.

B: It's not a coin flip but yeah.

E: There are patterns that can be determined sure.

C: Yeah and a social social psychologist, you know think think Kahneman think these different researchers who have you know kind of joined this idea of economics with psychology. Yeah we can predict behavior pretty pretty well. You're right some people are going to be on the tails of that normal curve but, yeah, but there's a normal curve.

S: Yeah, but there's a normal curve. We're not all over the place. Richard Wiseman talks about this all the time that's what he researches his interest is in looking at how to predict how people will behave with very specific stimuli.

E: Isn't that what like mentalism is practically about?

'S: Yes, yes absolutely.

C: For sure.

S: Mentalism wouldn't work if people were unpredictable, magic wouldn't work if people weren't predictable you know absolutely. So it's partly based on that premise, that people are going to be fairly predictable although this could theoretically also individualize the basic pattern you know, it's looking for your patterns not just the patterns of depression.

C: Right.

S: But then figuring out how to use, it's going to be the tricky part. And you know it could be they they talked about in the article recommending changes but that could be flipping the direction of cause and effect. It's like you're not sleeping as much and you're depressed so sleep more. Or it could just be that not sleeping is not causing you to be depressed it's a result of you being depressed. Or they're both the result of something else happening like increased stress and just saying sleep more is like a facelift it doesn't really treat the underlying problem which is changing a marker so this is where this is where the clinical research has to come in.

C: Yeah. I think in some ways that's one of the biggest problems with CBT, I think that you know we have made this the gold standard because it easily can be utilized in randomized controlled trials it lends itself to scientific. Like the way that we test drugs we apply to these psychological interventions and CBT just is good at that because it's highly manualizable, it's highly consistent, you can sort of do the same intervention to different groups that are randomized. The problem is that very often what we're looking at is a very acute intervention for a very acute symptom. And we're not really looking necessarily at why is that person having the panic attack. Like we might be able to intervene and teach them really good tools to sort of get grounded again, to get their breath back and to feel calm and feel safe and sort of not not go into the throes of panic. But what we're not getting deep into is why panic attacks consistently happen in their lives.

S: Yeah and that's I don't think this tool is for that you know.

C: That's what mental health treatment is for.

S: I agree, I think this will this is obviously not going to replace mental health treatment this is one more tool that could be useful for gathering behavioral information about specific individuals. And then also maybe providing some useful feedback either to them or to their practitioner. And you know you could you can envision a lot of situations where somebody who has obsessive-compulsive disorder they're starting to get into more and more obsessive behavior. Or someone who's anxious, who's starting to get physiological signs of anxiety that's starting to ramp up.

C: That they may not now this right away.

S: Oh yeah or somebody with depression or yeah leaving the house as much as they did yeah whatever. They might not be consciously aware of it yeah until it gets more severe and it might just trigger and sort of a more of an early warning system. So it'll be useful I think. But but at the same time I mean I think it will be used appropriately. But it will also then be completely abused it will be like─

C: Absolutely.

S: Download this app and treat your depression with this individualized personalized AI powered algorithm that will find out what what's making you depressed, you know what I mean. It'll be─

C: Yeah. Traditional therapy is hard and expensive but you can do this in only five minutes a day.

J: While you're sleeping.

E: So it'll unfairly target those who can't afford otherwise, to get the right kind of.

S: Well I don't know I don't know about that I, because this these kinds of things tend to target the worried well people who have disposable income. In fact there was a recent study that's that showed that the advent of tele-health you know with the pandemic, decreased socioeconomic disparities in access to healthcare.

E: Cool.

C: For sure it did.

S: So it's good it's been a very positive thing.

C: Yeah and I think an app could do that too.

S: Sure.

C: Because I think what a lot of people, what's important to realize if you've never gone through a heavy course of mental health treatment or sort of peeled the curtain back behind this whole industry, is that you know I had a supervisor who said it really well recently. She always tells her patients at the beginning my goal is to make myself...

S: Irrelevant, redundant.

C: Yes thank you not irrelevant or redundant but like my goal is to make─

S: Unnecessary.

C: ─unnecessary, my goal is to help you become your own therapist.

S: Yeah and that goes beyond mental health Cara because, especially with the advent of smartphones and technology you know a lot of modern medicine is teaching patients to take care of themselves.

C: Yeah it's like preventive care.

S: Yeah. Like yeah you can't be looking over your shoulder. You're gonna treat your own diabetes or migraines or whatever it is I'm just going to give you the tools to do it and teach you how to do it, but you've got to do it day-to-day because you're the one living with yourself you know I can't be there─

C: Absolutely.

S: To monitor what you're eating and you give you your insulin or whatever you got to do that. So this that's these kind of apps I think will help with that paradigm where we are really helping patients manage themselves. Because now this becomes an extra tool.

E: Patients generally do bad poor managers of their own when left to their own devices?

S: They're all over the place there but there are patients who are excellent, who are you know just like almost too much, like they're so completely on to you know on top of things and at the other end of the spectrum there are people like they don't even know what medications they're taking.

C: Yeah.

S: You will see the full full spectrum.

C: This tool could be used for very nefarious purposes or incredibly beneficial purposes.

S: Yeah and it will be used for both.

C: And it will be used for both yeah yeah.

S: Totally.

Hardy Tardigrades (30:17)

S: Yeah. The potential applications are huge. Of course, whenever you're talking about genetically modifying humans, that's a really high bar in terms of proving safety, but it's theoretically possible. And we may have to do it if we're really going to be a spacefaring race. But we'll see. Alright. Let's go on.

DNA Microfossils (43:29)

S: Have you guys ever heard of DNA microfossils?

J: No.

Belief in the Paranormal and Credibility (54:55)

Tesla Robots (1:09:17)

Who's That Noisy? (1:18:50)


New Noisy (1:21:44)

[_short_vague_description_of_Noisy]

really think about that one

Announcements (1:22:40)

Questions/Emails/Corrections/Follow-ups (1:23:48)

Email #1: Joe Rogan and Spotify

_consider_using_block_quotes_for_emails_read_aloud_in_this_segment_ with_reduced_spacing_for_long_chunks –

Science or Fiction (1:34:36)

Theme: Materials science

Item #1: For the first time scientists have been able to create a two-dimensional polymeric material, that is six times stronger than bullet-proof glass, and twice as strong as steel at one-sixth the density.[7]
Item #2: MIT engineers have created a biological “duct tape” that can be applied to internal organs, adhering in seconds and lasting for months.[8]
Item #3: Scientists have engineered a cultivar of cotton that creates fibers with almost the tensile strength of Kevlar.[9]

Answer Item
Fiction Kevlar-strong cotton
Science Super-strong 2D polymer
Science
Biological “duct tape”
Host Result
' win
Rogue Guess
Evan
Kevlar-strong cotton
Jay
Kevlar-strong cotton
Bob
Biological “duct tape”
Cara
Kevlar-strong cotton

Voice-over: It's time for Science or Fiction.


Evan's Response

Jay's Response

Bob's Response

Cara's Response

Steve Explains Item #1

Steve Explains Item #3

Steve Explains Item #2

Skeptical Quote of the Week (1:53:56)

Officials, educators, media personalities, and opinion makers must do everything they can to educate the public, assuage people’s fears, and show that the pandemic is being handled fairly and rationally. Lives and people's well-being are at stake, for years to come.
Nidhal Guessoum, Algerian astrophysicist

Signoff

S: —and until next week, this is your Skeptics' Guide to the Universe.

S: Skeptics' Guide to the Universe is produced by SGU Productions, dedicated to promoting science and critical thinking. For more information, visit us at theskepticsguide.org. Send your questions to info@theskepticsguide.org. And, if you would like to support the show and all the work that we do, go to patreon.com/SkepticsGuide and consider becoming a patron and becoming part of the SGU community. Our listeners and supporters are what make SGU possible.

[top]                        

Today I Learned

  • Fact/Description, possibly with an article reference[10]
  • Fact/Description
  • Fact/Description

Notes

References

Vocabulary


Navi-previous.png Back to top of page Navi-next.png